Structural reliability and risk analysis of elevated tanks.

摘要

When braced steel towers supporting elevated tanks are subjected to seismic loading, they usually exhibit a nonlinear hysteretic behavior. This behavioral pattern is instigated primarily due to the inelastic response of the bracing elements. Available nonlinear models, customized for stochastic analysis, apply only to symmetric hysteretic loops which do not accommodate the behavioral patterns exhibited by bracing elements when subjected to load reversals. In this study, a mathematical model that best describes the mechanical behavior of steel bracing elements is developed by adapting an existing model using logic algorithms to adjust itself to the mechanical behavior expected upon load reversal. The response statistics are then determined through the use of a Monte Carlo simulation technique. These response statistics are used to determine the reliability index of the structural system. The performance of the supporting tower systems is evaluated for X-shaped and Chevron bracing patterns, varying bracing cross sectional properties, as well as variable overall tower geometry. In addition, the modeling idealization is examined for two and three dimensional models. It was confirmed that the bracing elements have a major direct impact on the overall performance of the structure. However, the effect was found to be dependent upon the level of seismic energy input as well as the overall geometry and configuration of the structure. It was also established that the tower-tank systems are capable of undergoing hysteretic response cycles introduced through the damage sustained by the bracing elements. This behavioral trait served to enhance the tower's performance capacity against perpetual or periodic loading. Finally, the methodology developed herein, provides a more broad outlook on the behavior of these tower-supported tank systems because it utilizes a large number of samples of seismic input by means of which all various aspects of the system's response can be exposed and identified.